The poll results that poured in during the final week of politicking were puzzling.
Two had Baltimore City Mayor Martin O'Malley and Gov. Robert L. Ehrlich in a virtual dead heat in the gubernatorial race. Another had O'Malley leading with a 10-percentage point lead, others by just a few percentage points.
In the Senate race, one poll had Rep. Benjamin L. Cardin leading Lt. Gov. Michael S. Steele by a six-point edge. A few others had Cardin leading by three or five percentage points;, only one had him up by 10 points.
In the end, the winners were the same. Democrats O'Malley and Cardin won, by seven and 11 points, respectively, according to unofficial results.
Experts and pollsters attribute the largely unpredicted margin of victory of both candidates partially to a Democratic momentum that was building nationwide.
"There was an extraordinary coming together of the Democrat vote that was largely fueled by national conditions," said Keith Haller, president of the Bethesda-based Potomac Inc., which conducted a poll for The Sun that had O'Malley ahead by a single point. "It was a Democratic tsunami that basically carried across the state and caused people to vote Democratic in a way that we haven't probably seen in 10 or 20 years."
The disparities between some of the polls and the final outcome can be explained by a confluence of factors, pollsters say, the most prominent of which is predicting black voter turnout. In Maryland, unlike some other states, voter data is not broken down by race, making estimating those numbers somewhat of a guessing game.
According to exit poll results, the percentage of votes cast by black voters in the senate and gubernatorial race was 23 percent, a statistically high figure, experts said.
A Washington Post poll predicted that about 25 percent of likely voters would be black. That poll had O'Malley and Cardin leading by 10 percentage points.
The Potomac Inc. poll pegged the number at 19 percent. A Mason-Dixon Polling & Research Inc. poll used a 24 percent estimate.
"I think quite frankly the 10 or nine percent disparity is almost in itself attributable to the difference in the black voting numbers," said Patrick Gonzales, president of Gonzales Research in Annapolis, which did not conduct any of the polls.
Gonzales said predicting black turnout in Maryland is difficult because it's based largely on census data and historical estimates.
The numbers are crucial, he said, because black voters traditionally vote overwhelmingly Democrat, so a few percentage points can make a difference.
Steve Raabe, founder and president of OpinionWorks, an Annapolis-based opinion research firm, said the different numbers used for black voters could explain at least part of the difference in the polls.
In general, Raabe said, predicting voter behavior this year in Maryland was tougher than usual.
"It was a year where the turnout models that you might have used in the past really needed to be kind of reexamined because there were a lot of political forces buffeting people," said Raabe, who polled local races for The Capital newspaper in Annapolis.
"You had all the national unrest - the scandals in Washington, the bitter partisanship, and the aggressive campaigns being run in Maryland," said Raabe. "So there's a lot of different factors."